Laney p′ and u′ charts

attribute charts for very large samples

Limits = p̄ ± 3 · σz · √(p̄(1 − p̄)/n)

With very large samples the limits of a classic p or u chart become so narrow that almost every point falls outside, even in a stable process, because the real process varies more than the binomial or Poisson model allows (overdispersion). David Laney's charts widen the limits by σz, the moving-range spread of the standardised points.

σz close to 1 means the classic chart is fine; clearly above 1 with large samples calls for the Laney chart.

Calculate it

Attribute Control Charts: p, np, c and u charts for defectives and defects, including variable sample sizes.

Related terms

• p chart: control chart for the proportion defective